DE102016125953A1 - Method for operating a wind farm - Google Patents

Abstract

The invention relates to a method for operating a wind farm (112) having a plurality of wind turbines (100) or a wind power plant (100) for feeding electrical power into an electrical supply network (120), comprising the steps of: feeding electrical power into the electrical supply grid (120) Determining a forecasting power (P) that describes a power that can be fed in by the wind farm (112) or the wind energy plant (100) in a predefined future forecasting period, determining a reduction forecasting power (P) that designates a power for the forecasting period the power (P) fed in the forecast period can be reduced and offering the reduction forecast power (P) as a negative control power, the reduction forecast power (P) being determined as a function of the forecast power (P) and in dependence on a controllable consumption power, the controllable power consumption a performance bes which can be consumed on demand in the forecast period by controllable consumers (244) present in the wind farm (112) or in the wind turbine, respectively.

Description

The present invention relates to a method for operating a wind farm having multiple wind turbines. The present invention also relates to a corresponding wind farm.

Wind farms are known. They have several wind turbines and feed with these wind turbines via a common grid connection point in an electrical supply network. In addition to this main task of energy supply, such wind farms can also take on tasks for supporting the electrical supply network or at least make a contribution.

An important, at least possible contribution to network support is to reduce the power fed in if necessary. Wind turbines and thus wind farms have the property of being able to reduce their fed-in power very quickly. For example, you can halve your feed-in performance within a few seconds. Thus, wind turbines or wind farms are basically the producers in an electrical supply network, which can reduce their performance fastest without any problems. This means in particular that they can provide their power almost as quickly after the reduction again available. Especially for situations in which there is too much power in the relevant electrical supply network in the short term, wind energy plants can thus achieve rapid support by just as short a reduction in power.

In order to ensure a consistently stable electrical supply network, it is important for the operator of the electrical supply network, which is also referred to simply as a network operator, to have information about the performance that can be reduced. Wind energy plants or wind farms can basically reduce their fed-in power completely in a short time, which, however, presupposes that power is fed in as well. The power fed in depends in particular on the prevailing wind speed in a wind farm.

In order to allow the network operator, for example, to plan a day in advance, a wind forecast or weather forecast can be carried out in order to estimate the power that can be fed in and thus reduced for the forecast period.

Thus, such an expected performance can be estimated and the network operator can include this in his planning, but the performance is not known as sure that the network operator could count on it. For this reason, a kind of risk assessment is often made, which assumes that of a forecasted performance actually only a few percent, for example 3% can be assumed to be secure. So if a power of 1 MW is predicted, a safety consideration, which sets 3%, leads to a secured base value of 30 kW. Accordingly, only this security value is taken as a basis for remuneration for this system service. In other words, in the exemplary case mentioned, a negative one-minute power of 1 MW could be provided, whereas only 30 kW could be marketed.

The present invention is therefore based on the object to address at least one of the above-mentioned problems. In particular, a solution is to be proposed which makes it possible for a wind farm to be able to provide the highest possible negative control power, in particular minute control power, namely for a pre-planned period can safely provide so that the network operator can count on it, or the park operator market them to the network operator. At least should be proposed to previously known solutions an alternative solution.

According to the invention, a method according to claim 1 is proposed. The process is based on a wind farm that has several wind turbines. With these wind turbines electrical power is generated from wind and this electrical power is fed from this wind farm in an electrical supply network, namely usually via a grid connection point. Accordingly, therefore, electrical power is fed into the electrical supply network.

Basically, instead of a wind farm, only one wind energy plant can be used. The wind turbine can basically work like a wind farm, but usually provides less power than a wind farm and is therefore often not so relevant for grid support, but comes into consideration. All the following descriptions of a wind farm also apply mutatis mutandis to a wind turbine, as long as it does not peculiarities of the wind farm, which are not transferable to a wind turbine, such as features that affect the coordination of multiple wind turbines in a wind farm.

In addition, a forecasting performance is determined. The forecast power is a power that can be fed through the wind farm in a predetermined future forecast period. Especially here is a wind or weather forecast from which it can then be deduced how much power the wind farm can feed, for example, in the next 24 hours. In this case, it is also preferably taken into account whether all wind turbines are ready for operation or, for example, have to undergo maintenance, to name just one example of a boundary condition. The weather forecast can also take into account, for example, whether an ice situation is to be expected and accordingly some wind turbines can not be operated and / or only with simultaneous heating of the blades, to give a further example of a consideration.

Furthermore, a reduction forecasting performance is determined for the forecast period. This reduction forecasting performance is the power by which the power fed in over the forecast period can be reduced.

Then the reduction forecasting performance is offered as negative balancing power, in particular as a marketable negative balancing power. The wind farm transmits, for example, signals to a grid operator, which contain the information at what time or in which period the wind farm can reduce how much power its fed-in power.

It is now proposed that the reduction forecasting power is determined as a function of the forecasting power and as a function of a controllable consumption power. The determination of the forecasting power has already been explained, and the controllable consumption power is a power that can be consumed on demand in the forecast period by controllable consumers present in the wind farm. Such controllable consumers may for example be a blade heater, or a fan, to name but two examples. But there are also consumers who are not directly part of a wind turbine, but part of the wind farm. It is also possible to use electrical lines in the wind farm as electrical consumers. At least consumers are considered here who can be controlled on request. It is therefore about consumers who can be specifically targeted to a request, in particular to a request that generates a central control unit for consumption of power. It is particularly important that such consumers are driven in addition to and / or regardless of their usual purpose. The usual purpose of blade heating, to remain in this example, is to heat the rotor blade in question for deicing or for protection against icing. However, such a sheet heating can also be operated when no icing exists or is to be prevented. Insofar as this sheet heating mentioned by way of example can also be controlled independently of its original purpose of use, namely de-icing or anti-icing, in particular by a central parking control unit, it is a consumer controllable on request in the sense of the invention and can provide controllable consumption power.

To take another example already mentioned, namely transmission lines in the wind farm, so is their original purpose to transmit power, namely in particular of each one wind turbine to a grid connection point. If, in this case, reactive power is additionally transmitted here by a wind energy plant which is taken up by another wind power plant or to which reactive power that compensates accordingly by another wind energy plant, then this line is operated as a controllable consumer and is a controllable consumer within the meaning of the invention, if such consumption power can be controlled on demand, especially at the request of the central parking control unit. In particular, at least one wind turbine can generate such a reactive current that thereby reactive power is absorbed from the electrical supply network and leads to a power consumption in the electrical line.

It is now proposed to take into account for the entire wind farm, how much such or other controllable consumption power exists or can be activated. This total existing or activatable controllable consumption and the already explained forecast performance are now used as a basis to determine the reduction Progonoseleistung.

In particular, it was recognized here that the reduction prognosis performance can now be increased by this controllable consumption power in its amount. The forecasting performance may not be accurate or sufficiently certain, and thus may only be taken into account in part for the determination of the reduction forecasting performance, but at least it may be taken into account. Preferably, a minimum value is used by the forecasting power for a period of time within the forecasting period, so that the reduction forecasting power for the relevant forecasting period is calculated as a function of this minimum value.

In addition, for the determination of the reduction forecasting power, the controllable consumption power is added, which regularly also has the advantage that it can at least to some extent be regarded as secured. To stay with the mentioned example of sheet heating, the power consumption of this sheet heater known and it is also well known whether this sheet heater is otherwise needed in the forecast period or is available as a controllable consumption. For example, if it is enough, it is unlikely that this blade heater will be needed. Then it is preferably still considered whether depending on the outside temperature possibly only a partial heating is considered, in order to avoid overheating. Regularly, however, there may be the situation that the sheet heater can be operated permanently up to its nominal value. This nominal value is then available as controllable consumption power and can be used to determine the reduction forecasting power.

The reduction forecasting power can thus be composed of part of the forecasting power and the controllable consumption power. Namely, these values can be added up in terms of amount, ie the amount of the accountable part of the forecasting power and the amount of the controllable consumption power.

It is preferably proposed that the fed-in power be reduced to a request for network support in the forecast period by a value up to the amount of the reduction forecast power. It is therefore assumed that a reduction forecasting power was determined and transmitted to the network operator or his network control center as an offer, for example. If such a support case occurs, in which the grid operator resorts to this predicted and offered reduction forecast performance, the wind farm reduces its fed-in power accordingly. The network operator can then take part of the offered reduction forecasting service or he can take full advantage of this offered reduction forecasting service. The wind farm then reduces its fed-in power by the value that the grid operator uses.

However, the wind farm does not necessarily have to control the controllable consumers, but can first reduce the fed-in power. The fed-in power ideally corresponds to the previously forecasted performance, ie the forecast performance of the forecast period in which the wind farm is located. This forecasting performance has been taken into account, but not exclusively, in determining the reduction forecasting power offered to the network operator. In particular, it can be assumed that only part of the forecasting performance was taken into account when determining the reduction forecasting performance. It is therefore very likely that more power will be fed in than was assumed to be the guaranteed power fed in when determining the reduction forecasting power. Accordingly, the currently fed power can be reduced by a correspondingly more.

To give a simple example, a forecast power of 1 MW could have been predicted. Of this, 3% was taken into account as a guaranteed value for the determination of the reduction forecast performance, ie 30 kW, in order to stay with the example already mentioned above. In addition, it is assumed that a controllable consumption of 300 kW is present. The reduction forecasting power is then 330 kW for this example. However, if the forecasted power, ie the forecast power, actually sets in, namely at the level of 1 MW and the grid operator wants to resort to the complete reduction forecasting power, this means that a reduction of the fed-in power by 330 kW is required. The fed-in power can then be reduced to 670 kW, without even one of the controllable consumers being driven on demand, in order to reduce the controllable consumption power or a part thereof.

The decisive factor was that these exemplary 330 kW could not only be offered but also offered secured. For this purpose, the controllable consumption power was needed, so that the grid operator could rely on this 330 kW. If the actual support case then occurs, however, can only be considered a reduction of the actually fed power. Of course, however, the entire or a part of the then requested reduction performance can actually be met by the controllable consumers. Whether it is expedient to reduce the power generated or to remove part of the power generated by the controllable consumers can be decided individually and, for example, also depend on the type of consumer.

However, it is also the case that in fact significantly less power is generated and fed than was originally forecast as forecasting performance. If, for example, less power is generated and fed in as predicted and then a high reduction power is requested, it is also possible for power from the grid to be obtained from the wind farm. It is then fed so negative power.

So, for example, to stay in the above example, only a power of 200 kW fed and the network operator calls the fully offered reduction forecasting power as a reduction performance, so he asks for 330 kW reduction power, so the power generation can be reduced by the wind farm to zero , which reduces the input power by 200 kW becomes. In addition, 130 kW can be obtained from the grid and consumed by the controllable consumers. The wind farm then reduced its feed-in power by a total of 330 kW, ie as offered and, as requested by the grid operator, from 200 kW to 130 kW. This too is just one example that could be realized even with a single wind turbine. In fact, the method according to the invention can also be carried out with a single wind energy plant. Frequently, however, larger services than support services are desired by the network operator, so that then the use of a wind farm with multiple wind turbines is advantageous.

Preferably, it is thus proposed that one or more of the controllable consumers are turned on and / or controlled in their power loss, so that thereby the injected power is reduced, in addition power is taken from the electrical supply network and consumed when the fed, not yet reduced power is less than the reduction predictive power and is also less than a requested reduction power. If the fed, not yet reduced power is therefore less than the requested reduction power, the requested reduction performance is at least partially realized by the controllable consumers. It is assumed that the reduction performance is also requested only up to the reduction forecasting performance. Of course, it should not be ruled out to already set or reduce more reduction power than the reduction forecasting power. However, the determination and offering of the reduction forecasting performance is initially not based on this special case.

Preferably, the reduction forecasting power is determined to be composed of at least one of the production power or a value thereof multiplied by a predetermined quality value and the controllable consumption power. The forecasting performance is thus an achievement that is expected in the forecasting period and particularly, but not necessarily exclusively, depends on a wind and / or weather forecast. In order to derive a performance that is to be expected from this forecast performance, this forecast performance is multiplied by a predefinable quality value, which must therefore be <1 but> 0. For example, it can be 3%. The calculation can also be done for a period of time. If the forecast performance fluctuates over this period of time, a minimum value of the forecasting performance can be used.

Then there is the controllable consumption. This forecast power multiplied by the predefinable quality value or the forecast power value multiplied therewith and the controllable consumption power are added to the reduction forecast power. A corresponding example has already been given and explained above.

Preferably, the predicted availability of each controllable consumer is also taken into account for determining the controllable consumption power. Again, an explanation has already been given above for the example of sheet heating as a controllable consumer, according to which the operability of sheet heating depends firstly on whether it is even considered as a contribution to the reduction forecasting power or possibly already due to the weather according to their original purpose, namely the leaf deicing or anti-icing is used. In fact, if it is used for de-icing or anti-icing, its consumable power can not be part of the reduction forecasting performance, as it can not provide additional power in the event that feed-in power needs to be reduced. In this case, the predicted availability would be zero.

As another example it has been mentioned that the sheet heating can not be operated maximally due to very high temperatures in the environment because this could lead to high temperature values. If, for example, depending on the weather forecast, in particular the most unfavorable weather forecast to date, the sheet heating could only be operated at 50% of its rated output, the value 0.5 or 50% would be taken into account here as the forecast availability.

According to one embodiment, it is proposed that the controllable consumption power is composed of several partial services. Each partial power is assigned to a controllable consumer and thus describes the power to be taken into account by this controllable consumer. At least one of these partial services is calculated from a maximum output of their controllable consumer multiplied by a predicted availability of their taxable consumer. Preferably, several partial services are calculated in this way, in each case based on their controllable consumers and also in each case based on a predicted availability of their taxable consumer. Preferably, all partial services are calculated in this way.

For example, the taxable benefit can be the sum of two partial benefits. In this case, the first partial output could be assigned, for example, to the sheet heating already mentioned above. The maximum power of this consumer, so this sheet heater, the nominal power of the Be a consumer. For example, the sheet heater could have a rated power of 200 kW. If the weather forecast, as explained above by way of example, shows that only 50% of the sheet heater power can be used for the reduction forecasting power, the first partial power would be the product of 200 kW as the maximum power of the consumer, multiplied by 50 %, which is 0.5 as forecast availability. The first partial power would be 100 kW in this example.

The second sub-power could be assigned to a transmission line in the park, including a connected inverter that transmits or receives a reactive power over it. In this way, a maximum of 100 kW of power could be consumed and it could be consumed the less power, the more this power or at least one associated inverter for generating or transmitting power to be fed is needed. So if the forecast is that much wind is expected and a lot of power can be fed, the line can be used, for example, only to 30% still to power consumption. This would result in a value of 100 kW multiplied by 30% for the second partial power, so that 30 kW result.

In this last illustrative example of the second partial performance, however, instead of the predicted availability of 30%, the value 1, ie 100%, can be assumed, because the predicted availability of 30% results only when a lot of power is fed in. In turn, however, a lot of power can also be provided as a reduction power in case of cases where a power reduction has to be performed. The line as an additional consumer would only be needed if contrary to the wind forecast very little power would be fed, or even no power. In that case, however, the transmission line could also be used completely as a controllable consumer.

In the present example, the controllable consumption power would then result in 200 kW, namely from the first partial power to 100 kW and the second partial power likewise to 100 kW.

Preferably, it is proposed that the controllable consumers are controlled by a central parking control unit of the wind farm. About this central parking control unit of the wind farm, which can also be referred to simply as a wind farm controller, these controllable consumers of the wind farm can thus be controlled and thus controlled. In addition, they can also be taken into account as a whole for the park, and preferably the central parking control unit carries out the calculation or determination of the controllable consumption power in terms of amount. In particular, it also carries out the determination of the reduction forecasting performance. Thereby, the wind farm can act as a unit for such grid support by offering negative control power, namely by offering the reduction forecasting power, in or on the network.

In addition or alternatively, it is proposed that the central parking control unit receives a request signal from a network control center, which is provided for controlling the electrical supply network. This request signal may then form the request for network support. Furthermore, this central parking control unit preferably operates in such a way that it reduces the fed-in power as a function of this requirement and if necessary controls the controllable loads accordingly. As a result, the network control center, and thus in particular also the network operator who operates the electrical supply network, advantageously use a wind farm for grid support.

In particular, the network control center can activate a high negative control power in a very short time. At least part of this negative control power can be realized by driving consumers, which can also be called loads in this respect. It is possible to switch consumers on or off, or to control consumers at the level of their consumption. The output power of each wind turbine is controllable in height. In particular, by a combination of the control of the consumer and the control of the output of the wind turbines can be approached within technical limits, basically every operating point, the wind farm in its power output so can be adjusted continuously.

In accordance with a further embodiment, it is proposed that at least one energy store be present in the wind farm in order to provide, at least in the short term, additional power for wind power, which is generated by the wind energy installation of a wind farm, for feeding into the electrical supply grid. Such an energy storage, which may be a battery storage, for example, can thus feed additional power into the electrical supply network, at least he can provide power for it.

For this purpose, it is now proposed according to the stated embodiment that the reduction forecasting power is additionally determined as a function of available power of the at least one energy store. The reduction forecasting power can therefore also take account of power that this energy store has in the electrical supply network feeds in and thus the power fed into the electrical supply network can also be reduced. The consideration of such a performance of at least one energy storage device has the particular advantage that it can be safely predicted, because the memory content of this energy storage is likely to be known and its capacity to feed power into the electrical supply network, so which maximum discharge currents are allowed.

In addition or as an alternative, the determination of the reduction prediction power can also include such power that the energy store can absorb from the electrical supply network. This power is very well predictable for the energy storage and thus the offered reduction forecasting power can be increased by a correspondingly safe value.

Preferably, the reduction predictive power is used as the negative minute control power. The grid operator can therefore request from the respective wind farm that this reduces its fed-in power up to the level of the reduction forecasting power in the short term. This reduction can be requested and carried out within a few seconds. In particular, it is proposed that it be realized within a few seconds, especially within 30 s, preferably within 20 s and in particular within 10 s. Preferably, it is realized only for one or a few minutes, in particular less than 10 minutes. The power feed of the wind farm is thus reduced in a time <1 min and the reduction preferably lasts only a few minutes. Such network support can regularly be sufficient for short-term occurring in the electrical supply network transient processes for support. It is important that the network operator can be sure that such a reduction can be requested up to the amount of the reduction forecast service offered.

According to one embodiment, it is proposed that a central consumption control unit is provided for controlling or coordinating the activation of one, several or all of the controllable consumers. This central consumption control unit is preferably coupled to the central parking control unit and receives a total consumption target value from the central parking control unit. This total consumption target value indicates by what value the consumption of the controllable consumers should be reduced overall. The central parking control unit can thus specify such a total consumption target value and the concrete implementation or coordination is carried out by the consumption control unit. As a result, a structure can be simplified in particular, and the central parking control unit only needs to coordinate a reduction in the generated power with regard to the power fed in and the requested reduction power, which the network operator requests. It can reduce the power fed in and, if there is a need for further power reduction, can transfer that demand as the total consumption setpoint to the consumption control unit, which takes care of the coordination. Preferably, the consumption control unit is also used to calculate the total controllable consumption power. This value may be given to a prediction determination means, which may be part of the central parking control unit, so that this prognosis determination means or the central parking control unit can determine and offer the reduction prediction performance taking into account this controllable consumption performance.

Preferably, the forecasting power and, in addition or alternatively, the reduction forecasting power is specified as a time profile over the forecasting period. This allows temporal processes to be taken into account and stored. In particular, corresponding wind and / or weather forecasts can be taken into account, which regularly fluctuate over time.

Preferably, the forecast period is about 24 hours. This forecast period can also be set smaller or larger. It is preferably in the range of 12 to 48 hours, more preferably in the range of 18 to 36 hours. In particular, it is about a day. This will allow the network operator to plan for about one day. At the same time, this forecast period of about 1 day is a period for which wind and / or weather forecasting can be made with fairly good accuracy and reliability.

The prognosis period is preferably subdivided into 4-hour time segments, and a reduction prognosis performance is determined for each of these time segments. This can be offered for each of these periods, a reduction forecasting performance and in particular marketed. A network operator can thereby book such reduction forecasting services for 4-hour periods in order to have the possibility in the periods thus booked of requesting a negative control power if required.

Preferably, this prognosis period is divided into intervals, in particular at intervals of about 15 minutes, that is about four intervals per hour. Such a temporal subdivision into intervals creates the possibility of a temporal change of conditions over the forecast period well At the same time, however, stable values can be made available in the intervals. This can also be achieved that no large amount of data between wind farm and network operator or between wind farm and network control center must be transferred.

According to the invention, a wind farm is also proposed. This wind farm includes a feed-in device for feeding electrical power into the electrical supply network. Such a feed device may be provided in particular as an inverter at each of the wind turbines. In addition, a transformer may be provided at the grid connection point at or via which it is fed into the electrical supply grid.

In addition, this wind farm comprises a determination device for determining a forecasting power, which describes a power that can be fed in by the wind farm in a predefined, future forecasting period. The determining device may particularly include a process computer that is provided with appropriate information needed to determine the forecasting performance.

In addition, a prognosis determination means is provided, which is prepared for determining a reduction prognosis performance, which designates for the prognosis period a power by which the power fed in the prognosis period can be reduced. Such a prognosis determination means can also be designed as a process computer or comprise such. The prognosis determination means may also be part of the determination device or vice versa. The determination device and / or the prognosis determination means may be part of a central parking control unit.

In addition, a bidding means is proposed for offering the reduction forecasting performance as a negative control power. The offering means may also include a process computer and a communication interface for communicating with a network control center. This data can be exchanged between the network control center and the offering means. In particular, the offering means can transfer values to the network control center which specify the amount of reduction forecasting. Preferably, the communication is arranged to transmit such values at predetermined intervals, such as every 15 minutes.

The offering means may also be part of the determination device and / or the prognosis determination means and / or be part of the central parking control unit.

In addition, the wind farm has at least one controllable consumer who can consume power on demand. Preferably, several controllable consumers are provided.

For such a wind farm, it is now proposed that the determining device is set up so that the reduction forecasting power is determined as a function of the forecasting power and as a function of the controllable consumption power, in particular as described above. In this case, the controllable consumption power describes how, in the forecast period, power can be consumed on request by the controllable consumers present in the wind farm.

As a controllable consumption can also be considered an internal demand of the wind turbine, which is needed, for example, for fans or heaters or for lighting. Such a personal need can be for a modern wind turbine, for example, at 100 kW and have an availability of 100%.

Depending on the operating point, an additional domestic demand of 100 kW can be assumed, which can be switched on as a function of the operating point of the wind energy plant. Also, as a controllable consumer fan, heater or inverter cabinets, which are in a stand-by mode, concern: Since these consumers are used to a small extent according to their actual purpose and then not arbitrarily as a consumer to reduce the fed or fed Performance can be expected to be based on a reduced availability, which can still be 98%.

As further consumers, inverters or inverters can also be seen, which work as a feeding device. If they are underutilized, they consume less than at full capacity. This difference to full utilization can also be exploited as controllable consumption. For this purpose, the inverter or the inverter can be operated several wind turbines, for example, by mutual compensated reactive power supply, namely in a STATCOM operation. As a result, about 100 kW with an availability of 90% can be used per modern wind energy plant. Here then in each case the inverter and / or the line used is the controllable consumer.

Another consumer is the rotor blade heating to call, which can also be referred to simply as sheet heater and above as an example has been described. In a modern wind energy plant, depending on the operating point, about 500 kW with an availability of 98% can be used for each wind energy plant. Preferably, the wind farm is adapted to implement a method according to at least one embodiment described above. The wind farm preferably has a central parking control unit, which contains or communicates with the feed-in device, the determination device, the forecasting means and / or the provisioning means, or in particular has corresponding communication interfaces.

Preferably, a central consumption control unit is provided for controlling or coordinating the activation of one, several or all of the controllable consumers.

This central consumption control unit is preferably coupled to the central parking control unit and prepared to receive a total consumption target value from the central parking control unit indicating by what amount the consumption of the controllable consumers should be reduced overall. In particular, the central consumption control unit for this purpose has a communication means for communicating with the central parking control unit. Alternatively, the central consumption control unit is part of the central parking control unit.

According to one embodiment, at least one of the consumers is designed as an externally switchable consumer to be turned on by a control device in the wind farm for power consumption. Such a consumer can be referred to simplifying as a switchable consumer. In addition or as an alternative, at least one of the consumers is configured as an externally controllable load in its power consumption in order to be controlled by the control device in the wind farm for power consumption with a power that can be set in terms of height. Such a consumer can be referred to simply as an adjustable-height consumer. This means that the consumer can be switched or controlled externally, so to understand that the consumer is switched not only by himself or the device in which he is installed, or that he is driven by a device in the Wind farm is basically externally controlled and the control is so far switched outside the actual scope or controlled.

The control device may be the central parking control unit or the consumption control unit. For example, the central parking control unit can take over this switching and / or control of such consumers.

In order to precisely adjust a power reduction of the power fed in by the wind farm, the switchable consumers can be appropriately controlled by the height-controllable consumers and the wind turbines with respect to their generated power.

To give a simple example, if a wind farm generates and feeds 100kW of power and then has to reduce its injected power by 200kW, it can switch a switchable 150kW load and reduce its power by 50kW. The fed-in power is then reduced from + 100kW to -100kW, with the wind farm generating 50kW and the switchable consumer consuming 150kW.

The wind farm preferably has at least one energy store in order to provide, at least in the short term, additional power for wind power, which is generated by the wind energy installations of the wind farm, for feeding into the electrical supply grid. Such an energy storage device can also be contained in a wind energy plant, or in each case such an energy storage device can be provided in several or all wind energy installations of the wind farm. It is proposed that the at least one energy storage is prepared to reduce the additional power on demand by a network operator. In addition, it is prepared to provide information about its state of charge and / or a prediction of power reducible in the forecast period. In particular, he is prepared to transmit such information to the central parking control unit.

The at least one energy store can thus not only feed additional power, but this can also be reduced by request from network operators, so that in this way the energy storage can provide a negative control performance, in particular a negative minute control performance. For this purpose, he can also provide appropriate information about his state of charge and Leistungsreduzierfähigkeit. This makes it possible that this energy storage can also be included for the calculation of the reduction forecast performance.

For this purpose, the energy store has in particular a corresponding communication device in order to communicate with the relevant devices or means in the wind farm. In particular, it is proposed that such communication takes place with the central parking control unit, which transmits the relevant information about it and can process further. In the case of a reduction request by the network operator, this can be transmitted in particular from a network control center to the central parking control unit, wherein the central parking control unit can then transmit a corresponding reduction signal to the energy storage.

Preferably, the controllable consumers of the wind farm are also centrally controllable, namely in particular via the central parking control unit and / or via the central consumption control unit and / or via a central data network.

According to the invention, a wind energy plant according to claim 17 is also proposed. This works essentially or mutatis mutandis as described above to embodiments of a wind farm.

Preferably, the wind turbine is prepared to be used in a wind farm as described above in accordance with at least one embodiment and / or to carry out or participate in a method described according to at least one embodiment. Both the use in a wind farm, as well as the use in a described method means in particular that the wind turbine provides information about their operation and their consumers and / or that they and / or their controllable consumers for adjusting a power to be generated or a be switched off power or that any switchable consumers can be switched on and off, namely by an external control unit that is not part of the wind turbine or the consumer.

• 1 zeigt eine Windenergieanlage in einer perspektivischen Darstellung.
• 2 zeigt einen erfindungsgemäßen Windpark in einer schematischen Darstellung.
• 3 zeigt eine Struktur zur Bestimmung einer anbietbaren Reduktionsprognoseleistung.
• 4 zeigt eine Struktur zur Steuerung eines Windparks zur Bereitstellung negativer Regelleistung.
• 5 zeigt ein Leistungs-Zeit-Diagramm zur Erläuterung möglicher, eingespeister, verfügbarer oder angebotener Leistungen.
• 6 zeigt ein Leistungs-Zeit-Diagramm ähnlich der 5 aber mit einer Variation dazu.
• 7 zeigt ein weiteres Leistungs-Zeit-Diagramm mit einer weiteren erläuternden Variation gegenüber den 5 und 6.

The invention will now be explained in more detail below by means of embodiments by way of example with reference to the accompanying figures.

• 1 shows a wind turbine in a perspective view.
• 2 shows a wind farm according to the invention in a schematic representation.
• 3 shows a structure for determining an available reduction forecasting performance.
• 4 shows a structure for controlling a wind farm to provide negative control power.
• 5 shows a power-time diagram for explaining possible, fed, available or offered services.
• 6 shows a performance time diagram similar to the 5 but with a variation to it.
• 7 shows another power-time diagram with a further illustrative variation over the 5 and 6 ,

1 shows a wind turbine 100 with a tower 102 and a gondola 104 , At the gondola 104 is a rotor 106 with three rotor blades 108 and a spinner 110 arranged. The rotor 106 is set in operation by the wind in a rotary motion and thereby drives a generator in the nacelle 104 at.

2 shows a wind farm 112 with exemplary three wind turbines 100 that can be the same or different. The three wind turbines 100 are thus representative of basically any number of wind turbines in a wind farm 112 , The wind turbines 100 put their power, namely in particular the electricity generated via an electric parking network 114 ready. In this case, the currents or outputs of the individual wind turbines are generated 100 and mostly is a transformer 116 provided that transforms the voltage in the park, then at the feed point 118 , also commonly referred to as PCC, into the utility grid 120 feed. 2 is only a simplified representation of a wind farm 112 which, for example, only shows three wind turbines, although often more than three wind turbines are present in a wind farm. Also, for example, the parking network 114 be designed differently, in which, for example, a transformer at the output of each wind turbine 100 is present, to name just another embodiment.

The wind farm 112 of the 2 feeds by means of several feed devices 230 into the electrical supply network 120 one. The feeding devices 230 are part of the wind turbines 100 , Each feed device 230 For example, it may be a frequency inverter that generates a corresponding three-phase current across the transformer 116 then into the electrical supply network 120 is fed. For the overall control of the wind farm 112 is a central parking control unit 232 intended. Furthermore, a detection device 234 provided for determining a prognosis performance, which with a prognosis determination means 236 is coupled. The forecasting means determines a reduction forecasting performance and this can be determined by the offering means 238 be offered as negative balancing power. The terms reduction forecasting performance and negative balancing power can also be used synonymously. The investigative device 234 , the prognosis determination agent 236 and the tender 238 can also be combined into a single entity, which is in the 2 is indicated.

Information on the wind turbines serves as the basis for determining the forecasting performance as well as for determining the reduction forecasting power 100 over the data network 240 be transferred, namely in particular to the determining device 234 and above that to the forecasting agent 236 , The data determined or determined thereby can be sent to the central parking control unit 232 for example, from the offering agent 238 be transferred from. The investigative device 234 and / or the forecasting title 236 and / or the offering 238 can also be part of the central parking control unit 232 be.

In particular, the particular reduction forecasting power may then be provided by the central park control unit 232 a network control center 242 Tobe offered. The network control center 242 may then request a negative control power even in case of cases where appropriate network support appears necessary. This can also be done via communication with the central parking control unit 232 respectively. The central parking control unit 232 is in this case provided so that it can then control the wind energy installations 100 correspondingly or predefine corresponding setpoint values. For this she can also use the data network 240 serve. In addition, it is considered that the negative control power is also made by appropriate control of consumers in the wind farm 112. Such consumers can be part of every wind turbine 100 be such as fans in the wind turbine 100 that are known as such and therefore in 2 not shown separately. However, external consumers are also considered, for whom an external consumer is involved 244 is indicated. The control and in particular also the coordination of the consumers to reduce power and thereby provide negative control power is performed by a central consumption control unit 246 , This central consumption control unit 246 can also do this via the central parking control unit 232 be controlled and the central consumption control unit 246 can control the consumer, including the external consumer 244 over the data network 240 make.

3 In the structure shown, it particularly illustrates the prognosis determination means 236 of the 2 , here as a forecast determination block 336 is shown. The forecast determination block 336 receives as input variables from a weather forecast, in particular the predicted temperature values or the predicted temperature profile, as T _W (t) in the prognosis determination block 336 received. The temperature can influence, for example, the expected operation of the wind energy plant. In the worst case, too low a temperature can lead to the shutdown of the system. In other cases, the system is operated reduced or it is operated a heater.

As a further value, depending on a wind forecast, which may be part of the weather forecast, a predicted wind power or a course of the predicted wind power, which here as P _{W, P} (t) is shown.

Furthermore, a prognosis uncertainty p _P (t) is received. This forecast uncertainty depends in particular on the quality and the boundary conditions of a weather forecast and can be specified as a probability distribution. It often results in a value of 3%. This means that if 3% of the forecasted performance is offered as a negative balancing power every quarter of an hour, this can be achieved with a probability of 99.98% of the quarter-hourly values. The value of 3% can be a current value, but this can be increased especially if predictions especially of the wind improve, so that a forecast uncertainty is provided here as an input variable. Furthermore, for each consumer in the wind farm, the consumption power P _{L, n} that characterizes it is received. The characteristic power P _{L, n} can be especially its nominal power and / or its maximum power. This value is a characteristic quantity and basically a fixed value. The 3 shows for the one input P _{L, n} , but which is only representative of the corresponding performance of each consumer. This is indicated by the index n. So there is such a value for every consumer considered.

Furthermore, for each consumer, an availability value V _{L, n is entered} as a relative value, especially as a percentage value. This indicates for each consumer whether and in what amount this is available. The value can range from 0 to 100%, with 0% indicating that the consumer is unavailable, whereas 100% indicates that the consumer is available in full of his distinctive performance P _{L, n} .

These values can then be used in the forecasting block 336 a maximum allowable power is determined, namely in the sense that this maximum adjustable power can be provided with a security of 99.98%. Such a calculation takes place for a period or a time block of 4 hours. The result is then a time-dependent maximum controllable power P _neg.min (t), which corresponds to the reduction _forecasting power or can be synonymously referred to as reduction _forecasting power. Maximum adjustable power here means that can be regulated up to this value. The network operator, to whom this maximum adjustable power P _neg.min (t) is offered, can therefore adjust up to this value or demand a reduction. However, due to safety considerations, and especially considering forecast uncertainties, this is the least amount of power that can be offered. So, for example, with better forecasting or when lower security is accepted, even more control power could be offered.

This maximum adjustable power P _neg.min (t) is then the offer block 338 which can then offer this control power to the network operator or can market it for network support. In that regard, the offer block concerns 338 which could also be called a marketing block, the marketing of the negative balancing power.

4 explains how to coordinate compliance with a requested negative control power when such a negative control power is actually requested. This is the coordination block 450 intended. This receives as inputs the power value P _M , which indicates how much power the wind farm can generate and feed in total, and the value P _A , which indicates how much power the wind farm is to be regulated, by how much power the wind farm so less should feed. Both are values that relate to the wind farm as a whole.

The coordination block 450 then calculates a power P _{W, n} to be generated and fed by it for each wind energy plant, and a power P _{L, n} to be consumed by consumption at the respective plant. This will be the coordination block 450 as far as possible an optimal combination of generated power and consumed power. These results, ie the power P _{W, n} to be generated and fed by each installation and the power P _{L, n} to be consumed by each wind turbine _, are then sent to the respective wind energy installations 400 transferred and representative is only one wind turbine block 400 shown. In fact, but go into the coordination block 450 two sizes and it will output 2 • n sizes again.

5 shows a power-time diagram in which represents the wind farm generated and feedable power P _WP as a function of time. In this respect, the diagram shows the power P _WP for the entire wind farm under consideration. On the ordinate is a negative branch also indicated an available consumption power P _V, but the first in the modified chart 6 and 7 becomes relevant. The present diagram, the same applies to the diagrams of 6 and 7 , shows a time frame over four hours and thus over a period over which a negative control performance could be offered in principle. Basically, however, for a longer period, for example, for 24 hours, a forecast and also a forecast performance and a reduction forecasting performance can be specified. However, such a prognosis period would preferably be divided into several 4-hour sections. In the case of 24 hours, this area would be divided into six areas of four hours each.

The present diagram shows a possibility for a prediction power P _{W, P} with the dotted graph. This forecast power P _{W, P} depends on a weather forecast, especially from a wind forecast. At time t ₁ , the forecasting power P _{W, P} assumes a minimum value, which is plotted as P _{W, P, min} . The calculation of a marketable negative balancing power is based on this minimum value of the forecasting performance. Such a marketable negative control power P _{nR is} calculated, for example, as 3% of the minimum forecast power P _{W, P, min} . The marketable negative control power thus also corresponds to the reduction forecasting power 5 the power P _M actually available for injection is plotted as a solid line. The marketable negative control power P _nR is illustrated here as a shift, namely, lowering of the power P actually available to _M shown in phantom reduced power fed P _Mr. Correspondingly, the distance between the actually available power P _M and the power P _Mr injected in the case of a grid support request is the marketable negative control power P _nR . In the example shown, this distance should have 3% of the minimum forecast power P _{W, P, min} and is shown enlarged only for the purpose of better illustration in this and also the next representation.

The 5 shows by way of illustration the special case that the actual available power P _M at time t ₂ has fallen so far that it only has the value of the marketable negative control power P _nR . In fact, this negative control power P _{nR is chosen} as the described 3% of the minimum forecast power P _{W, P, min to} preclude with a certainty of 99.98% that the actual available power P _M drops lower in a quarter-hour interval. Usually, therefore, the actually available power P _{M will} fall only in 0.02% of the 15-minute intervals in the year so far and also the reduced feed-in power P _Mr will come in only 0.02% of quarter-hours in the vicinity of the zero line.

In any case, the distance or the difference of this minimum value of the available power P _M to the minimum forecasting power P _{W, P, min to be} regarded as a forecast error F _P. Because a large forecast error can occur, the marketable negative control power P _{nR is} calculated in a conservative _manner . To do this 5 illustrate that in spite of such a non-predicted strong drop of the power P actually available _M, the reduced power P _Mr nevertheless does not become negative. That is the reason why the already explained conservative factor of 3% is frequently chosen.

For illustration, a fake feed power P _{Mf is} drawn. This fictitious feed-in power P _Mf is a shift or decrease of the actually available power P _M by the full value of the minimum forecast power P _{W, P, min} . This illustrates the case when, for the same situation, as the marketable negative control power, the minimum forecast power P _{W, P, min} would have been used directly, ie not 3% thereof, but 100%. In fact, in the event of such a case, if such a large control power were to be demanded, the reduced feed-in power would take this value of the fictitious feed-in power P _Mf . It can easily be seen that the power to be fed in reduced would become negative and thus could not be realized by reducing the power fed in, because a fed-in power can only be reduced to the value zero.

Accordingly, this negative power or the negative portion of the power is not marketable and this value is _plotted as the distance to the zero axis to illustrate the fictitious feed-in power P _Mf at the time t ₂ . This non-marketable negative _balancing power is shown as P _nRn .

To solve the invention has now been recognized and proposed to take into account consumption of consumers in the wind farm. Such a consumption power P _LS , which is certainly available, is in the 6 otherwise the 5 corresponds, drawn. This safely available consumption power P _LS is shown here for the sake of simplicity as a horizontal line. In fact, however, the safely available consumption power can also fluctuate, for reasons which have already been explained above. However, the fluctuation is likely to be low in the case of the certainly available consumption power P _LS compared to the forecast power P _{W, P.} Alternatively, it is also possible here to assume a constant value for the safely available consumption power P _LS , namely the value which is smallest in magnitude for the considered 4-hour time period, if fluctuations are assumed.

7 now explains how this safely available consumption performance can improve the marketable negative control performance, namely increase the amount. This is illustrated by the fact that the reduced, feed-in, instantaneous power P _Mr can be shifted by this value of the certainly available consumption power P _LS , as a result such that the thus-changed, reduced, feed-in power P _{Mr '} the power consumption that is just certainly available P _LS touched. This results in a significantly increased marketable negative control power P _{n R ',} that of the negative control power P _{neg, min} (t) of the image of the structure 3 equivalent. This is in the 7 indicated, but without the temporal dependency. The time dependency is to be understood as being different for different 4-hour periods. 7 but shows only a 4-hour cut and for this marketable negative control power P _{neg, min is set} as a constant value.

As a result, the non _- marketable negative _regulatory power P _{nRn 'is} smaller in magnitude than in the case of 5 ,

It was recognized that the actual feed-in may be lower than the forecast feed-in due to forecast errors. The same applies to the available negative control power because it depends on the feed-in.

The proposed solution is particularly important for such a wind farm to be able to provide control power.

For the prequalification of power plants, including wind farms, for the provision of control power, the available and thus spontaneously controllable power must be demonstrated. This often requires a security of 99% or even 99.98%. It must therefore be proven regularly with a reliability of 99% or 99.98% that the offered control power can also be delivered. Such a standard benefit may be offered and remunerated, the remuneration being determined by the offered standard benefit. It depends not on the retrieved, but on the offered control power.

With today's forecast quality, only approx. 3% of the forecast feed-in power can also be offered as a control power, which results from the consideration of the probability of forecast errors.

It has been recognized that by operating internal consumers of a wind turbine or wind farm, the amount of control power to be offered can be significantly increased.

In one example, instead of -126kW, calculated as 3% of rated power (-4,200kW * 0.03) -892kW, it can be sold, with an additional 766kW of taxable consumers (-4,200kW * 0, 03-766kW).

For this purpose, it is proposed that a first technical device determines the minimally secured control power in the corresponding time slots, taking into account the weather, the wind forecast and the loads.

Another technical device controls the individual plants and consumers in the wind farm to realize the control power requirement. By a method for combining discrete consumers and controllable feed, it is possible to follow the required load curve, so to realize.

Claims (18)

A method of operating a wind farm (112) having multiple wind turbines (100) or a wind turbine (100) for feeding electrical power into an electrical utility grid (120), comprising the steps of: - feeding electrical power into the electrical grid (120), - determining a forecasting power (P _{W, P} ) which describes a power which can be _{fed in} by the wind farm (112) or the wind energy plant (100) in a predefined future forecasting period, - determining a reduction forecasting power (P _nR ) which designates a power for the forecasting period by which the power (P _M ) fed in the forecast period can be reduced, and - offering the reduction forecast power (P _nR ) as a negative control power, wherein - the reduction forecast power (P _nR ) - depending on the forecast power (P _M ) and - depending on a controllable consumption, is determined, with the controllable consumption a Le describes can be consumed in the forecast period by existing in the wind farm (112) or in the wind turbine, controllable consumer (244) on request. Method according to Claim 1 characterized in that the injected power (P _M ) is reduced to a request for grid support in the forecast period by a value up to the amount of the reduction _{prediction power} (P _nR ). Method according to Claim 1 or 2 , characterized in that one or more of the controllable consumers (244) are turned on and / or controlled in their power decrease, thereby the fed power (P _M ) is reduced, in addition to power from the electrical supply network (120) and taken is consumed when the injected, not yet reduced power (P _M ) is less than the reduction _{predictive power} (P _nR ) and less than a requested reduction power. Method according to one of the preceding claims, characterized in that the reduction _{forecasting power} (P _nR ) is determined so as to additively composed at least of - the production power (P _{W, P} ) or a value thereof, in particular a minimum value thereof, multiplied by a predeterminable quality value, and - the controllable consumption power, wherein in each case a predicted availability of the respective controllable consumer (244) is preferably taken into account for determining the controllable consumption power. Method according to one of the preceding claims, characterized in that the controllable consumption power - is composed of several partial services, - each partial power is assigned to a controllable consumer (244) and - one, several or all of the partial services each - from a maximum power of its taxable consumer (244), multiplied by a predicted availability of its taxable consumer (244). Method according to one of the preceding claims, characterized in that the controllable consumers (244) are controlled by a central parking control unit (232) of the wind farm (112) and / or the central parking control unit (232) by a Netzleitstelle (242), which for controlling of the electrical supply network (120) is provided, receives a request signal as a or the request for network support and reduces the injected power (P _M ) as a function of this requirement and possibly controls the controllable loads (244) accordingly. Method according to one of the preceding claims, characterized in that - at least one energy store in the wind farm (112) or the wind energy plant (100) is present, at least in the short term additional power to the wind power generated by the wind turbines (100) of the wind farm (112) or the wind power plant (100) is generated to provide for feeding into the electrical supply network (120), wherein - the reduction _forecasting power (P _nR ) is additionally determined in dependence on available power of the at least one energy storage _device . Method according to one of the preceding claims, characterized in that the reduction _{forecasting power} (P _nR ) is used as a negative minute control power. Method according to one of the preceding claims, characterized in that for driving or coordinating the activation of one, several or all of the controllable consumers (244) a central consumption control unit (246) is provided, wherein the central consumption control unit (246) preferably with one or the central park control unit (232) and receives a total consumption target value from the central parking control unit (232) indicating by what amount the consumption of the controllable consumers (244) is to be reduced overall. Method according to one of the preceding claims, characterized in that the forecasting power (P _{W, P} ) and / or the reduction forecasting power (P _nR ) are specified as a time profile over the forecasting period. Method according to one of the preceding claims, characterized in that - the forecast period is 12 to 48 hours, preferably 18 to 36 hours, in particular about 24 hours and / or - the forecast period is divided into 4-hour periods and / or - the forecast period is divided into 15-minute intervals. A wind farm (112) comprising a plurality of wind turbines (100) for feeding electrical power into an electrical utility grid (120), comprising: - a feeder (230) for feeding electrical power to the electrical utility grid (120), - a determiner (234) for determining a forecasting power (P _{W, P} ) describing a power that can be _{fed in} by the wind farm (112) in a predetermined future forecasting period, - a forecasting determination means (236) for determining a power forecasting power (P _nR ) designating a power for the forecasting period, by which the power fed in during the forecast period can be reduced, and - a bidding means (238) for offering the reduction forecasting power (P _nR ) as a negative control power, - at least one controllable consumer (244) which can consume power on request, the - determining device ( 234) is arranged to - that the reduction forecasting power (P _nR ) is determined as a function of the forecasting power (P _{W, P} ) and, depending on a controllable consumption power, the controllable consumption power describing a power which in the forecast period is determined by the at least one controllable load (112) present in the wind farm (12). 244) can be consumed on request. Wind farm (112) after Claim 12 , characterized in that it is prepared to perform a method according to any one of Claims 1 to 11 in particular, that it has a central parking control unit (232) and that the central parking control unit (232) executes the method in whole or in part. Wind farm after Claim 12 or 13 , characterized in that for driving or coordinating the activation of one, several or all of the controllable consumers (244) a central consumption control unit (246) is provided, wherein the central consumption control unit (246) is preferably coupled to one or the central parking control unit (232) is and is prepared to receive a total consumption target value from the central parking control unit (232) indicating by what amount the consumption of the controllable consumers should be reduced overall. Wind farm after one of the Claims 12 to 14 characterized in that - at least one of the consumers is configured as a consumer switchable from an external source in order to be switched on by a control device in the wind farm for power consumption, and / or - at least one of the consumers is designed as an externally controllable consumer in its power consumption, to be controlled by the control device in the wind farm, the power consumption with an adjustable power in height, and wherein the control device is preferably designed as one or the central parking control unit and / or as one or the consumption control unit. Wind farm (112) after one of the Claims 12 to 15 , characterized in that - at least one energy storage in the wind farm (112) is present to at least short-term additional power to wind power, which is generated by the wind turbines (100) of the wind farm (112) to feed into the electrical supply network (120) wherein the at least one energy store is prepared to - reduce the additional power on demand by a network operator and - provide information about its state of charge and / or prediction of power reducible in the forecast period, in particular to the central parking control unit (232) transfer. A wind turbine (100) for feeding electric power into an electric utility network (120), comprising: - a feed device (230) for feeding electrical power into the electrical supply network (120), - a determination device (234) for determining a forecast power (P _{W, P} ), which describes a power that can be fed in by the wind turbine (112) in a predefined future forecasting period, - a forecasting means (236) for determining a reduction forecasting power (P _nR ), which designates a power for the forecasting period by which the power fed in during the forecasting period Power can be reduced, and - an offering means (238) for offering the reduction _{forecasting performance} (P _nR ) as a negative control power, - at least one controllable consumer (244) _{capable of consuming} power on demand, wherein - the determining device (234) is adapted to - that the reduction _{forecasting power} (P _nR ) depending on the forecasting power (P _{W, P} ) and - depending on a controllable consumption power, is determined, the controllable consumption power describes a performance in the forecast period by the at least one controllable load (244) present in the wind turbine (100) on request can be consumed. Wind energy plant (100), in particular after Claim 17 , prepared to be in a wind farm (112) after one of the Claims 12 to 16 to be used and / or a method according to one of Claims 1 to 11 to perform or participate in it.

Images